Hugging Face
Models
Datasets
Spaces
Buckets
new
Docs
Enterprise
Pricing
Website
Tasks
HuggingChat
Collections
Languages
Organizations
Community
Blog
Posts
Daily Papers
Hardware
Learn
Discord
Forum
GitHub
Solutions
Team & Enterprise
Hugging Face PRO
Enterprise Support
Inference Providers
Inference Endpoints
Storage Buckets
Log In
Sign Up
axjns
/
strix-halo-kernels
like
1
kernel
rocm
amd
triton
strix-halo
gfx1151
rmsnorm
geglu
flash-attention
attention
License:
apache-2.0
Model card
Files
Files and versions
xet
Community
Copy to bucket
new
main
strix-halo-kernels
35.2 kB
Ctrl+K
Ctrl+K
1 contributor
History:
4 commits
axjns
Fix stale tuning claim; add attention tags and probe docs
0f129c5
verified
3 days ago
build
Add Triton flash attention (2-3x over AOTriton, 31x less peak memory)
3 days ago
.gitattributes
Safe
1.52 kB
initial commit
3 days ago
README.md
7.1 kB
Fix stale tuning claim; add attention tags and probe docs
3 days ago
sdpa_probe.py
2.44 kB
Add Triton flash attention (2-3x over AOTriton, 31x less peak memory)
3 days ago
test_bench.py
3.89 kB
Fused RMSNorm + GEGLU/SwiGLU Triton kernels for gfx1151
3 days ago
test_flash.py
4 kB
Add Triton flash attention (2-3x over AOTriton, 31x less peak memory)
3 days ago
